1 J. Hey- Genetic Species

The species problem is the persistent species. biological and philosophical debate on the meaning of the word "species" and the methods of species identification. With a meaning of "species" that follows from a simple model of DNA replication, species are shown to be real and non-arbitrary groups of organisms, under some circumstances. However, it also follows that many organisms do not belong to species. The criteria by which a group of organisms can be considered a species is whether they share in a process of genetic drift. This simplification is a negative resolution to the problem cases of species identification; it permits a concise listing of the causes of diversity and of the reasons why species can be very difficult to identify, but it does not simplify the process of species identification. For population biologists, a reduced species concept reveals a research plan for the study of organismic diversity that focuses on the determinants of structure in patterns of genetic drift. The finding that species exist, but that some organisms do not occur in species, reveals the central difficulty of systematic theories that assume the existence of species. The diversity of life seems to have a pattern whereby organisms fall into a limited number of types. Although the existence of these types, or species, has long been recognized (Mayr, 1982) , the definition of the word "species" and the identification of species that the word "species", as often used by biologists, signifies a distinct kind of biological individuality (akin to "organism" or "cell"), and does not simply reject species as a distinct kind of individual (Nelson, 1989) , and among those who do not, there persists a lack of consensus on the defining properties of this kind of individual (Endler, 1989). Two questions remain much discussed: is it useful to consider species as individuals?; and if so, what is the defining attribute shown that these relatively modest phenomena will create a kind of individual that has a close correspondence with other concepts of the meaning of The concept that is developed (the genetic species) is similar to some elements of the cohesion species concept (Templeton, 1989, 1994). In particular, both species concepts rely extensively on the idea of shared genetic drift. However, the two concepts differ in their motivation and their purpose. Templeton begins his discussion with the question ?What is a species?" and the implicit assumption that species as individuals exist in nature. The genetic species concept arises from the basic question of whether organisms actually occur in groups that are individuals (i.e. species). A SIMPLE SYSTEM Consider the reproduction of a single-celled asexual organism. To simplify, focus solely on the replication of the DNA genome, and view the remainder of the organism as the machinery of DNA replication. This simplification follows from the genotype/phenotype relationship: the genotype is the information for the organism and is replicated; while the phenotype is the organism and is recreated each generation as a function of the genotype. The focus is on the transmission of information, and much of this discussion should apply in principle to any informational replicating system (Orgel, 1992). Hereafter, "a DNA" will be used to refer explicitly to the molecule that contains the genotype information and that is replicated in this simple system. This choice of term is motivated by the same J. Hey Genetic Species 2 Figure 1 The gene tree history of a sample of DNAs. The tree is a hypothetical depiction of a true history, not to be taken as an example of a tree estimated from data. Key features include: the directionality of time, from the past to the present; branches; branch tips; and nodes, the junctions of branches. Branch tips refer to different pieces of DNA that exist at the present moment. The remainder of the diagram below the tips is a description of history. The tip of the branch at the base of the tree is undefined because the true history is not known beyond this point. Branches refer precisely to the persistence of a DNA sequence through time. This persistence means at times the physical persistence, but also includes numerous cases of replication when it is the information in the sequence that persists. The nodes of the tree refer precisely to those cases of DNA replication when both daughter sequences that were produced are ancestors of sequences that are represented as tips of branches. reasoning that lead Dawkins to employ “replicator”, context, "homologous" means that the different DNAs which he defined as an “entity that interacts with its are related by common ancestry and thus share a gene world, including other replicators, in such a way that tree history (Fig. 1). A sample of DNAs may include copies of itself are made” (Dawkins, 1978). The one or multiple sequences. word “gene” is avoided Consider a DNA that undergoes replication to here, and by Dawkins (1978), because of its form two daughter DNAs, and suppose that the conventional meaning as a single unit of function in replication depends upon both the DNA sequence and the expression of the phenotype. In this paper, a DNA the local environmental resources. After replication, is a contiguous double stranded molecule that is a the fates of the daughter DNAs may be linked because replicator. A DNA may be physically connected to a they coexist under common circumstances and longer contiguous stretch of DNA or it may correspond compete for the same pool of resources. If resources to a single chromosome, depending on the context. A are limiting and competition occurs so that not all DNA sequence is a particular order of the component DNAs undergo replication, and if both daughter DNAs nucleotides within a DNA and is not synonymous with and all of their descendants are subject to the same a DNA. The terminology also includes "DNAs" to circumstances (i.e. no mutational differences or refer to multiple pieces of homologous DNA. In this geographical separations), then the long term persistence of both groups of descendants is mutually exclusive. After some time, perhaps after many rounds of replication, one group of descendants will have replaced the other, or both will have been replaced by the descendants of yet another DNA that also shares those circumstances. Now consider that DNAs reside within organisms, and that the continuous random replacement of DNAs by the descendants of others is caused by a random birth and death process that happens within a group of organisms that share a finite set of resources. With an allele-based model of genetic variation, the effects of the random birth and death process within a population of organisms include random changes in allele frequencies that lead to the random loss and fixation of mutations. This process is called genetic drift. From a gene tree standpoint, genetic drift is manifested as a randomly shifting pattern of coancestry among a set of DNAs (Fig. 2). Consider the persistence through time of a group of organisms that experience a random process of birth and death, and then consider the gene tree, one DNA per organism, with the tips constantly moving forward with time. The random death of some organisms means that some gene tree tips do not persist, and the branches that lead to these tips disappear from the gene tree history that remains for those DNAs that do persist and replicate (Fig. 2). This shift forward in time of the pattern of ancestry proceeds continuously, and at intervals will include forward jumps for the most basal node representing the ancestor for an entire group of DNAs (Watterson, 1982). Two kinds of events can cause the descendants of two daughter DNAs to not be mutually exclusive. First, the daughters may differ because of mutation, and this may cause differences in the circumstances of replication. Individuals carrying the mutation may utilize resources in a different way so that they do not compete directly with individuals not carrying the mutation. Second, one daughter DNA and respective J. Hey Genetic Species 3 Figure 2 A gene tree at successive times under genetic drift. Asterisks at branch tips at times A and B indicate sequences that did not persist to the next time period. Solid lines at times B and C indicate branches that were present at the previous time and still remain (and are now longer) in the tree because of the persistence of DNAs at the branch tips. Dotted lines at times B and C indicate new branches leading to DNAs that arose by replication since the previous time period. All solid lines at times B and C correspond to a line (solid or dotted) at the previous time, and all lines at times A and B (solid or dotted) that do not lead to an asterisk, correspond to a solid line at the next time. descendants may occur in a geographically distinct of life, though the actual nucleic acid may have been location from other DNAs. Under both mutation and single stranded RNA (Gilbert, 1986). The model is geographic separation, the genetic drift experienced by also an approximation, for the multicellular case, of the the descendants of one DNA occurs partially transition from the reproductive cells of an organism in independently of that experienced by the descendants one generation to the reproductive cells in a of the other. To describe this in another way, the descendant organism in the next generation. For a individual DNAs within a group compete more directly group of multicellular organisms, the appropriate gene with one another, and are more likely to be replaced by tree history to consider is one in which a single DNA the descendants of other DNAs within the same group has been taken from each organism. In this case many than by the descendants of DNAs from the other but not all of the instances of DNA replication group. In this way both mutation and geographic represented by branches on a gene tree will have separation can lead to multiple groups of DNAs that occurred during germ line development and somatic are not mutually exclusive. growth (to the degree that somatic growth occurs prior The model of replication that leads to multiple to germ line development). The remaining replication groups of DNAs that are not mutually exclusive has events along gene tree branches, and all those three components: a DNA with a sequence that causes replications represented by gene tree nodes, must have replication; the possibility of mutations; and some kind occurred within reproductive cells that gave rise to of environmental structure such that the pool of gametes or offspring. resources used by one group of DNAs need not GENETIC SPECIES completely overlap those of another group of DNAs. The basic criteria by which something, a group This simple system probably existed early in the origin of organisms for example, is considered to be an individual is whether or not the location of components are constrained in space and time (Ghiselin, 1966, 1974; Hull, 1976, 1978). In contrast the members of a class have no spatiotemporal constraints. If a species is a class, then the constituent organisms are members that simply happen to share the particular features by which the class is defined. If a species is an individual then the constituents are components that collectively form a distinct entity (Ghiselin, 1987), and thus has a real existence independent of an observer. It is proposed that organisms whose DNAs share in a process of genetic drift constitute a kind of biological individual. A process of shared genetic drift among DNAs will occur within a group of organisms that share a birth and death process. A shared “birth and death process" means that at any point in time individual organisms can be physically replaced, and their function in the environment can be replaced, by the descendants of other organisms that are within the group. It also means that organisms compete, so that the survival and reproductive success of one organism has an affect on the survival and reproductive success of other organisms. This type of competition and potential for replacement has also been described as demographic exchangeability (Templeton, 1989) and occurs when organisms share the same niche, in an ecological sense (Hutchinson, 1958). The idea that organisms within a species share a niche is also a component of several other species concepts, including the evolutionary species concept (Simpson, 1961), as well as a recent formulation of the biological species concept (Mayr, 1982 p. 273), and it is contained within the ecological species concept (Van Valen, 1976). A similar view, emphasizing the competition J. Hey Genetic Species 4 that occurs when organisms closely share resources, has been described by (Ghiselin, 1974) in an exposition on species as individuals (as opposed to classes). Ghiselin proposed that "species" be defined as "the most extensive units in the natural economy offspring. (Dobzhansky, 1950) such that reproductive competition occurs among their parts". "Genetic species" will be used to refer to a group of organisms that share genetic drift. This term has been chosen because "genetic" is often used in reference to the material of inheritance and because genetic species arise from the process of replication, which is the essential function of DNA. "Genetic" has also been used in the context of species concepts in reference to the process of gene exchange (Simpson, 1961; Masters and Spencer, 1989). The present meaning of "genetic species" may be appropriate, despite the limited prior use of "genetic", because replication is a more fundamental function of DNA than is recombination. Genetic drift has long been recognized as a force that contributes both to uniformity within populations and to variation among separate populations (Gulick, 1872, 1888; Wright, 1931; Templeton, 1989). Genetic drift is proposed not as a mechanism that causes species, but as a description of the process of species existence. Genetic drift is the name of the instantaneous process (like a mathematical derivative taken with respect to time) of a genetic species. In this view, the ultimate cause of species is the reproduction and death of organisms that compete for shared resources. One could also say that the organisms in a genetic species share in a common birth and death process. However, organisms reproduce because of instructions in the genotype, and genetic species existed prior to the existence of complex cellular phenotypes (i.e. organisms). THE ROLE OF SEX In a gene tree view of the history of a sample of DNAs, sex is synonymous with recombination, and can be defined as any process that causes different portions along the sequence of a set of DNAs to have different gene tree histories. In the absence of sex, the gene tree history of a sample of DNAs is the same for all parts of the sequence. With sex, it is possible that the speed of genetic drift for one portion of a DNA is different from another portion. If there is a high recombination rate, then a large sample of DNAs will have a history of many different gene trees, perhaps as many as there are base pairs in the sequence. For sexual organisms, a genetic species is the same as a Mendelian population, as defined by Dobzhansky: A Mendelian population is a reproductive community of sexual and cross-fertilizing individuals which share in a common gene pool. . . .The smallest Mendelian populations are panmictic units (Wright, 1943), which are groups of individuals any two of which have equal probability of mating and producing Thus by definition, organisms within a Mendelian population share in a probabilistic process of reproduction, and all pairs of organisms are equally subject to reproductive failure and equally likely to reproduce. Within a Mendelian population, each generation occurs with some distribution of reproductive success among the component organisms. The shape of this distribution may vary across generations, but at any point in time the particular pattern of reproduction is a major determinant of the gene tree for all portions of the genome. A sample of DNAs for a short region of the genome will have a particular history, while a different genomic region will have a different history; yet all of these histories must run through the same historical procession of organisms, with a different group of reproductives each generation. Thus a Mendelian population carries genomes with numerous gene trees that were all shaped by a common birth and death process. GENETIC DRIFT AND NATURAL SELECTION From a genetic perspective, natural selection can be defined as variation in reproductive success caused by genotypic variation (Lewontin, 1970), and it is often cast as a directed force of evolutionary change in contrast to the random force of genetic drift. However at the level of DNA where there is linkage, natural selection on functional DNA sequence variation contributes to the genetic drift that occurs among linked sequences. In a genetic species of asexual organisms, a mutation that changes a DNA sequence and causes natural selection, also causes a new pattern of genetic drift among organisms that carry that mutation. In effect, a new genetic species is created by the mutation; although one of the species will probably be replaced by the other. For the DNAs of organisms with recombination, the acceleration of genetic drift by natural selection depends on the degree of linkage, the number of sites of functional variation, and the strength of natural selection on the functional variation (Hill and Robertson, 1966; Felsenstein, 1974). Natural selection on functional genotypic variation may play a major role in the formation of new genetic species. However, shared genetic drift, and not natural selection, is the appropriate description of the essence of genetic species. Genetic species will share in the process of natural selection on functional DNA sequence variation, and thus will share adaptations. However this process proceeds both concomitantly with, and as a contributor to, genetic drift. Furthermore, from a genealogical perspective J. Hey Genetic Species 5 (Fig. 2), genetic drift proceeds even in the absence of polyploid or parthenogenetic. In either case, the DNA sequence variation and in the absence of natural progeny of the hybrids can no longer exchange genes selection caused by DNA sequence variation. with the original species and do not share in a common THE CAUSES OF MULTIPLE SPECIES process of genetic drift with either original species. There are two kinds of events that can cause a The description of the causes of multiple species single genetic species to become two genetic species. is intended as a simplification. By posing the First is physical distance or the emergence of a discussion in terms of gene tree histories and genetic physical barrier between organisms so that they do not drift, mutations and barriers to the movement of DNAs draw from the same pool of resources. This appear to be the only possible causes of species geographic barrier to drift may be reversed if diversity. Similarly the irreversible aspect of mutation, organisms are mobile or if geography changes. as a cause of diversity, contrasts with the reversible Second, is the appearance of a mutation that changes effect of barriers, and this distinction follows directly the sequence of a DNA so that an organism and its from the genetic drift perspective. However, these descendants undergo genetic drift separately from mechanisms are not novel ideas, but rather a simple other organisms not carrying the mutation. This kind version of the causes of speciation that have been of speciation can be reversed only if all of the discussed in other contexts. In particular, geographic descendants of the organism fail to reproduce so that isolation and the evolution of barriers to sex form the all copies of the new sequence ceased to exist. Back central model of speciation under the biological mutation could not reverse the speciation event unless species concept (Mayr, 1942; Dobzhansky, 1951). all descendant copies of the mutant DNA underwent Also, a similar depiction of the impact of different back mutation. kinds of mutations, and their shifting effect as a The effects of geographic barriers and function of sex, is contained within the cohesion environmental heterogeneity on genetic drift do not species concept (Templeton, 1989). change as a function of sex. If geography constrains GENETIC DRIFT AND POPULATION the replacement of some individuals by the STRUCTURE descendants of others, then it also constrains the For a group of organisms to be a genetic process of recombination between some pairs of species, genetic drift must be creating a kind of individuals. This constraint on genetic drift will occur individual, meaning an entity with boundaries in space for all portions of the genome regardless of sex. and time (Hull, 1976). Perhaps the clearest example of The kinds of mutations that can create genetic a genetic species would be a group of organisms that species are different for sexual and asexual organisms. are completely panmictic (i.e. random mating) amongst In the absence of sex, a new advantageous mutation themselves without the occurrence of any mating with causes an organism and its descendants to undergo a organisms outside the group. For asexual organisms, different pattern of genetic drift from those organisms the criteria of completely random mating is replaced by not carrying the advantageous mutation. Within a one of complete demographic exchangeability genetic species of sexual organisms, favorable (Templeton, 1989). In an asexual group of organisms mutations do not contribute directly to the with complete demographic exchangeability, any one multiplication of species. Regions of the genome organism’s physical place and environmental role under tight linkage to the site of the mutation will could be taken by any other organism. However, experience accelerated genetic drift and have a complete panmixia or demographic exchangeability shortened gene tree history (Kaplan et al., 1989) while need not be present in order for there to be a sharp the gene trees of unlinked portions of the genome will boundary to the pattern of genetic drift. In general, not be affected. In short, sex prevents favorable any time that a group of organisms occurs with gene mutations from causing one genetic species to split flow or demographic exchangeability among the into two organisms that lack the beneficial mutation organisms, and where the level of gene flow or are not excluded from the birth and death process that demographic exchangeability is high relative to the occurs among organisms that carry the mutation. level with organisms outside of the group, a boundary However, there is a class of mutation that, in obligately will exist (Templeton, 1989). sexual organisms, can contribute to the formation of It is also possible for there to be variation in the new genetic species. These are mutations that cause degree to which organisms share genetic drift, and this recombination either not to occur between some variation may not occur with sharp boundaries. individuals or cause the results of recombination to fail Consider the case of isolation by distance (Wright, to reproduce. Included within this class of events are 1943) in which the times of possible coancestry for a genomic changes that shift the mode of reproduction or given pair of DNAs are proportional to the physical the ploidy level of the genome. For example, distance between the members of that pair. Under this interspecific hybrids of sexual organisms may be scenario the pattern of genetic drift, as well as the J. Hey Genetic Species 6 pattern of genetic variation, among organisms may not Dobzhansky’s portrayal of nested levels of gene flow, be structured but may follow a continuous pattern over beginning with high gene flow and panmixia at the some environmental landscape. A sample of DNAs lower limit, does not necessarily imply the existence of will still have a gene tree history, but the spacing of the a sharp boundary at the upper limit where gene flow nodes may vary widely and is expected to have a approaches zero. Under the biological species concept, variance larger than expected under a simple the existence of a sharp boundary to gene flow is demographic model of shared genetic drift (Slatkin, attributed to the presence of isolating mechanisms that 1987; Hey, 1991). prevent gene flow. The genetic species concept Another kind of population structure may lead differs from the biological species concept in not to nested levels of demographic exchangeability or having a necessary role for any process other than gene flow, with multiple nested boundaries to the genetic drift. pattern of genetic drift. An example is the population A focus on genetic drift as the essence of structure of Escherichia coli. Genetic drift may occur species provides a form of negative answer to some over a short time scale among the cells in a single colony on a petri dish, and over a longer time scale among the population of cells within the intestine of a single mammal. On a larger scale, exchange of E. coli cells occurs between different individual intestinal populations, and this leads to turnover of individual intestinal populations (Hartl and Dykhuizen, 1984). Thus the structure of E. coli populations seems to include a hierarchy of levels of genetic drift. Individuality, by the criterion of a boundary in the pattern of genetic drift, occurs at multiple nested levels. Another pattern not clearly consistent with shared genetic drift can arise when two sexual populations share genetic drift for just a portion of the genome. They may share drift over the entire genome with the exception of a single region that is under natural selection, with different functional forms of the region maintained in different populations. Similarly, genetic drift may be shared for organelle genomes but not nuclear genomes. At the other extreme are two populations that share drift over very little of the genome because of natural selection. It is possible that these populations may still generate hybrids with some reproductive success and share drift at parts of the genome that are not linked to those that are under differential natural selection. These scenarios of population structure and hybridization illuminate an area of uncertainty for the biological species concept (Mayr, 1963) (ch. 2). This can be seen by considering the very close parallel between Dobzhansky’s (1950) concept of a Mendelian population, and the genetic species concept. Dobzhansky defined the smallest Mendelian population as a panmictic unit, and envisioned larger Mendelian populations to be groups of panmictic units that engaged in gene flow. Finally, the largest Mendelian populations are biological species. Dobzhansky envisioned the existence of a boundary, a partition in the magnitude of gene flow such that there was a point when species could be defined. However, the concept of a Mendelian population does not by itself imply the existence of such a boundary. aspects of the "species problem". All three of the situations described (complex population structure, isolation by distance, and sexual populations with some genomic partitioning of gene flow) are "problem cases" for which biologists are often at a loss for clear ways to delineate species. A positive resolution of these uncertainties would be, for example, some description of the meaning of ?species" that permitted objective resolution of many problem cases. However, when these problem cases are considered with the reduced concept of genetic species, the uncertainty of these situations does not go away. In short, it appears that species are not a necessary consequence of those processes that do often give rise to species. A similar negative resolution was also proposed by Leven (1979) for many of the problem cases that occur among plants, especially those that rarely outcross or have limited gamete dispersal and experience isolation by distance. In these cases, and others where organisms do not occur in groups that share genetic drift, organisms do not occur as parts of species. In the long term, it is expected that all organisms have histories that include periods when ancestors were part of a genetic species. This is because the causes of genetic species, both mutational and environmental, will sometimes create groups of organisms with periods of uniform genetic drift in which the probability of recent coancestry for any pair of DNAs has little variation. A low variance among pairs of DNAs for possible coancestry times is more likely for a group of DNAs among which genetic drift is proceeding rapidly. Rapid genetic drift may result either from ecological circumstances that sharply curtail reproduction or from the appearance of a strongly favored mutation. Also, with sex, rapid genetic drift can occur for a tightly linked portion of the genome as a result of advantageous mutations or an abundance of deleterious mutations (Maynard Smith and Haigh, 1974; Kaplan et al., 1989; Charlesworth et al., 1993). Thus environmental changes or mutations may create genetic species from groups of organisms that are not in genetic species. J. Hey Genetic Species 7 CONTEMPORANEOUS SPECIES AND HISTORY species concepts includes this very exercise for the One of the major dichotomies that arises in the case of contemporaneous populations of interbreeding discourse on species concepts is between organisms (Hennig, 1979 p.73; de Queiroz and contemporaneous, or “snapshot” viewpoints and Donoghue, 1988; Nixon and Wheeler, 1990; Davis and historical viewpoints (Endler, 1989). A Nixon, 1992; Baum and Shaw, 1995). This can also contemporaneous view is most likely to be useful to a be done for the genetic species concept. The historical population geneticist or ecologist focusing on ongoing extensions of contemporaneous situations, with and population processes. The biological species concept, without sex, can be considered. and the genetic species concept are contemporaneous 1) For organisms that do not have sex, the concepts. However, systematists take a historical view common ancestors of a genetic species become and generally refer to species within the context of fewer, the further one goes back into the past, ancestor-descendant relationships. For example, and this reaches a limit of a single individual. (Simpson, 1961) defined a species as an “ancestralFor two genetic species that have recently descendant sequence of populations”. Cracraft defines diverged, there may be many shared ancestral a phylogenetic species as “an irreducible (basal) cluster organisms, and these may or may not have of organisms, diagnosably distinct from other such existed as a single genetic species. Thus clusters, and within which there is a parental pattern of whether or not a historical viewpoint includes ancestry and descent” (Cracraft, 1989). One kind of ancestral and descendant species depends on resolution of these two different viewpoints of species how much time is being considered and the (contemporaneous and historical) has been to argue history of genetic drift. For an ancestor and that both are real but distinct kinds of entities descendant that are far apart in time, the (Donoghue, 1985; de Queiroz and Donoghue, 1988). ancestor is literally a single organism and could In particular, de Queiroz and Donoghue (1988) argue not be a species. Furthermore, this is true that there are two processes, interbreeding and whether or not the descendant organisms are a common descent, that are both valid species criteria. contemporaneous genetic species. However, the two criteria differ fundamentally in their 2) For a group of sexual organisms, different incorporation of time. Interbreeding is a process that portions of the genome will have different gene can be viewed in a short time interval (e.g. a tree histories. The nodes of these gene trees are generation). Common descent is an explicitly historical ancestral portions of the genome, and the spatial criterion whether or not a group of organisms have and temporal distribution of the organisms that descent in common depends on what ancestors they carried these ancestral DNAs can be considered had. The processes of interbreeding and common as a function of the historical pattern of genetic descent are distinct merely because one (i.e. common drift. From models of the variance of the descent) includes the passage of time. If there exists a coalescent process in the presence of population kind of species that can be defined by historical structure, it is clear that the variance of the time relationships, then these historical relationships must and the geographic location of gene tree nodes have arisen because of processes that occurred over can become large, arbitrarily large, depending time. For example, Simpson’s definition of an on the degree to which populations of ancestral evolutionary species explicitly refers to the existence organisms depart from panmixia (Slatkin, 1987; of a population that exists at a particular point in time. Hey, 1991). It is possible for a group of Cracraft’s definition of a phylogenetic species refers organisms to exist as a contemporaneous genetic explicitly to a “parental pattern of ancestry and species, and yet have a history of ancestors that descent”. Though not defined precisely, this definition did not occur in genetic species. This type of clearly implies some contemporaneous process by history would cause different gene tree which groups of parents leave offspring. In both of estimates, for different portions of the genome, these cases, Simpson’s and Cracraft’s, the definition of to have a very large variance for the pattern of a historical species concept supposes the existence of node spacing (see EMPIRICAL some kind of cohesive group of organisms that exists CONSIDERATIONS). It is also possible for a for each slice of time in the history of the historical group of organisms that do not exist as a species. contemporaneous genetic species to have a The gap between historical and “snapshot” history of ancestors that did occur as genetic concepts can be bridged by considering the process species. that creates a contemporaneous species, and then extending that process through time to generate a In summary, when the genetic species concept is picture of a historical species. Indeed much of the extended to a historical context, the exercise does not literature that forms the debate on the phylogenetic reveal the emergence of a distinct historical entity. J. Hey Genetic Species 8 There is no epiphenomenon that could be called a drift based on patterns of genetic variation must admit historical species when the history has not included the two kinds of uncertainty. First, it is possible that a persistence of contemporaneous genetic species. In group of organisms that seem to have a recent history the literature on phylogenetic systematics, one of genetic drift, may not currently share genetic drift. graphical tool is to depict historical species as a tube, Secondly, a group of organisms may currently share with time as an axis that runs the length of the tube. In genetic drift, but this may be due to a recent mixture of these diagrams, a view of a species or a population at a historically separated organisms that did not occur in a point in time can be represented as a cross section of a single species. The current situation may be one tube (e.g. (Hennig, 1979 p. 59). The point made here genetic species, but the recent history which is is that the histories of organisms need not have a shape reflected in the pattern of genetic variation, may that can be represented in this way. In the absence of include zero or multiple genetic species. sex, the more recent ancestors of a group of organisms It is not the purpose of this report to develop may not have shared genetic drift. For times longer in practical criteria for the identification and delineation the past, the ancestor of a group of asexual organisms, of genetic species. Rather the purpose is to show that regardless of whether they occurred as part of a genetic an assessment of genetic drift is required if species are species, is a single individual. In the presence of sex, to be identified and distinguished. In this view, the the ancestors of a contemporaneous genetic species task of identifying species and understanding the may not have been a genetic species. The ancestors details of the causes of speciation falls squarely within may have been spread across a wide expanse of the domain of population genetics. Since the time of geography, in an isolation by distance relationship, or (Wright, 1931) much of the field of population with a complex structure of multiple populations. genetics has consisted of research on ways to assess EMPIRICAL CONSIDERATIONS genetic drift, and on the effects of genetic drift. The Biologists often face the question of whether a point that a population genetic approach must be used sample of organisms comes from one or more than one to identify species and understand the causes of species species. Under the genetic species concept it is also has been repeatedly emphasized by Templeton (1981, possible that some or all of the organisms are not part 1989, 1994). of any species. In practice, assessments of genetic drift It is useful to provide a qualitative description of are expected to fall into two different categories, some criteria that can be used for the case of a set of instantaneous and recent. An instantaneous DNAs, one from each member in a sample of determination is an assessment based on ongoing organisms. It is possible to describe the kinds of gene patterns of reproduction in contemporaneous tree histories that can occur for a set of DNAs under organisms. For obligately sexual organisms, an different evolutionary models (Hudson, 1990) and it is assessment of genetic species status would be the same possible to estimate the true tree for a sample of DNAs as an assessment of Mendelian population status if there is some variation in their sequences. Although (Dobzhansky, 1950). Those organisms that are DNA sequences and gene tree estimates are not the exchanging genes are necessarily also sharing in a only way to study genetic drift, they are increasingly process of genetic drift. In this case, an assessment of referred to in the context of the population genetic genetic species status is reduced to determining if the causes of speciation (Hey, 1994; Templeton, 1994; pattern of genetic drift has sharp partitions such that Baum and Shaw, 1995). there are distinct groups having gene flow within and Consider a sample of two sets of homologous little gene flow between. For organisms that are not DNAs, one from each of several organisms from each exchanging genes, an instantaneous assessment of drift of two candidate species and consider the null model could be made on ecological and demographic grounds that the sample comes from a single species by an assessment of demographic exchangeability. (Templeton, 1994). The gene tree history of the entire However, this is tantamount to measuring the sample will, in the absence of recombination, be fundamental niche for each of the organisms in the representable as a bifurcating diagram (Fig. 1). If the sample (Templeton, 1989) which is generally ancestral DNAs collectively underwent genetic drift, impractical. An alternative to instantaneous measures then the times between successive nodes of the tree are are assessments of recent patterns of genetic drift based also a function of a genetic drift process. For several on patterns of genetic variation. With electrophoretic quite simple demographic and linkage/selection data on protein variation or with comparative DNA models the distribution of times between nodes has sequence data, the patterns of variation can be been solved (Tavaré, 1984; Hudson and Kaplan, 1988; interpreted in terms of the genetic drift that has Takahata, 1988). These theoretical compositions are occurred in the time since the variation arose. called "coalescent" models (Kingman, 1982), However, the genetic species concept is a reflecting the pattern of a collapsing sample as one contemporaneous one, and so an assessment of genetic proceeds from the present into the past. The general J. Hey Genetic Species 9 Figure 3 Three hypothetical gene trees. In each case, four DNAs have been randomly sampled from each of two genetic species, Y and Z. Species Y is larger with a slower rate of drift and larger intervals between nodes than species Z. See text for further explanation. prediction is that the most recent nodes of a tree will be particular common ancestor (Hennig, 1979 p. 73). more closely spaced in time than the more distant Figure 3C shows relatively little partitioning of the nodes, and that the expected time between successive gene tree with respect to sample. This pattern could nodes is proportional to the population size (Hudson, occur if the ancestors of one group recently became 1990). However the details of this prediction vary geographically separated from others of the same considerably depending on the demographic model genetic species, so that relatively little drift has (Tajima, 1989). occurred within each group. If a sample includes DNAs from two groups of If genetic drift proceeds among the descendants organisms with a collective history of a single genetic of groups Y and Z, then the gene trees depicted for drift process followed by a shift (speciation) to two group Y in Fig. 3B and groups Y and Z in Fig. 3C will drift processes, then the structure of the gene tree will eventually be replaced by monophyletic gene trees (see differ in two general ways from the tree for a sample e.g. (Avise and Ball, 1990)). The emergence of from a single genetic species. First, the distribution of monophyletic gene trees is caused by the forward shift times between successive nodes after the speciation of the pattern of ancestry that occurs within a group of will be the result of two independent systems of DNAs that share genetic drift (Fig. 2). genetic drift. In the terms of a coalescent model, the The condition of having previously identified waiting time between successive nodes will be a candidate species (as in Fig. 3) is useful for articulating function of two random variables instead of a single various patterns of genetic drift. However, in practice, one. Second, the tree will be partitioned into sections researchers must consider cases where multiple species that are exclusive of DNAs from one of the sampled may exist, but where the DNAs have not been labeled groups (Baum and Shaw, 1995). Figure 3 shows three beforehand as belonging to candidate species. This examples of multiple DNAs from each of two uncertainty adds considerably to the difficulty of independently drifting populations, one with a faster identifying species simply because a posteriori rate of drift (i.e. short time intervals between nodes) and one with a slower rate of drift. In Fig. 3A, the two samples form separate subtrees; while in Fig. 3B, one sample forms a gene tree that is nested within the gene tree for the other sample. Groups Y and Z of Fig. 3A and group Z in Fig. 3B are monophyletic, meaning that the group includes all of the descendants of a hypotheses have much stricter criteria of statistical acceptance than a priori hypotheses. For sexual species, this burden can be reduced by generating hypotheses regarding genetic species using data from one locus. A second, unlinked, locus can then be studied to test these hypotheses, which are now a priori (Hey and Kliman, 1993; Hey, 1994). For sexual organisms in a genetic species, different parts of the genome will have different gene tree histories, though nearly all portions are expected to share genetic drift. The actual rate of drift will vary among genomic regions by chance and because natural selection and variation in recombination rates will cause the rates of genetic drift to vary across the genome. One kind of natural selection that can frustrate a gene tree assessment of species status for sexual organisms, and thus require the study of multiple portions of the genome, is balancing selection. This type of natural selection occurs when there exists a stable pattern of multiple sequences, or alleles, for some region of the genome. The persistence of multiple functional forms will create gene trees like those in Fig. 3A, in this case with designations Y and Z referring to different alleles. Genetic drift will occur within each allele class, but natural selection prevents the replacement of one allele class by descendants of the other. PROPOSALS FOR POPULATION BIOLOGISTS Biologists often apply "species" without a clear meaning of the word or, if a specific meaning is articulated, with uncertainty over whether a group of organisms actually fit the meaning. For cases when "species" is used to convey some degree of J. Hey Genetic Species 10 individuality on the part of a group of organisms, as parts of species, may be incorrect. For example, meaning some degree of spatio/temporal integrity, a investigators in the field of phylogenetic systematics two step convention is proposed. First, the specific strive to use historical relationships among organisms meaning that should be used is that of genetic species. as a guide for the classification of organisms (Hennig, However, for most groups of organisms, including 1979 p.73). It is implicit, and sometimes explicit, many that might be called species under other within this perspective that organisms truly do occur in concepts, the population structure will not resemble phyla (e.g. species or higher taxa). For instance panmixia. Thus the second component of species Hennig stated that species and higher taxa “are all identification is the inclusion of a description of the segments of the temporal stream of successive pattern of genetic drift. ‘interbreeding populations’” (1979 p. 81). If Having a meaning of "species" makes it possible interbreeding populations are not, in fact, a necessary for population biologists to avoid the word and occurrence for organisms, then the theory of associated uncertainties. If genetic drift underlies phylogenetic systematics has an error. However, as mechanistic species concepts, it follows that an severe as the theoretical implications may be, the assessment of the history of genetic drift for some practical implications are largely unknown. It is not group of organisms will obviate the need for species known how many organisms do not occur in genetic identification of those organisms. One or more species species, and it is not known to what degree the names may be applied, but will convey no additional ancestors of present day organisms occurred in genetic information. species. Also, the estimation of gene trees as a way to GENETIC SPECIES AND SYSTEMATICS estimate phylogenetic relationships will appear to be It has been proposed that some organisms do insensitive to whether or not ancestral organisms were exist in groups that are real, and that the in species. For instance, in the absence of sex there spatiotemporal distinctness of these groups arises when will still exist a bifurcating gene tree history for extant organisms share in a process of genetic drift. The DNAs, regardless of the species status of ancestors (see genetic species is also a distinct kind of individual, CONTEMPORANEOUS SPECIES AND HISTORY). different from the individual status that might be Such a tree may be misidentified as a phylogeny, when considered for larger groups of organisms. For it actually does not represent phyla. example, it has been argued that species are not If the theoretical implications of the genetic different from genera or other taxa, and that a taxon at species concept were to be included in a theory of any level can be thought of as an evolutionary unit systematics, then at least two possible courses can be (Nelson, 1989). However, the criteria of genetic considered. First, a theory of systematics could focus species is shared genetic drift among organisms. It is solely on the historical relationships of individual possible that some organisms occur in a pattern of organisms (Vrana and Wheeler, 1992). This method population structure with nested levels of genetic drift would have the advantage that the identification of among organisms (see GENETIC SPECIES AND individual organism is vastly easier than for individual POPULATION STRUCTURE) so that genetic species species. However, such a system would also face the may occur within larger genetic species. However this difficulty that for any point in time, a sexual organism pattern is still caused by genetic drift among will likely have many ancestral organisms. This means organisms, which is in turn caused by gene flow and that a bifurcating hierarchical tree will often not be a demographic exchangeability. These causes of individuality at the species level need not be the same as processes that occur among species. While it may be argued that there are processes of species turnover that are analogous to genetic drift, these processes will not be identical to gene flow and demographic exchangeability. In short, the genetic species is a distinct kind of individual that includes multiple organisms. The finding that an organism may not be part of a species, either in a contemporaneous sense or a historical sense, has implications for the study of the historical relationships among organisms and for classifying organisms. These implications may be roughly categorized as theoretical and practical. The primary theoretical implication is that a systematic theory that assumes that all organisms occur in nature good historical model. An alternative proposal is that a systematic investigation could begin with a process of identifying candidate species and estimates of historical relationships among candidate species. For example, candidate species could be identified by relatively practical criteria (e.g. morphological similarity). This general approach is commonly used with other species concepts. In addition systematics would include a process of examining the genetic species status of candidate species. This kind of systematic protocol, with a theoretical component that includes the genetic species concept, would also need to address the issue of the classification of organisms that are not in species. Both of these proposals, using organisms as the individuals of systematics and including the genetic species concept within systematic theory, address the J. Hey Genetic Species 11 theoretical basis of systematics. A third alternative species concepts are examples. In contrast, would be to forego biological theory, and develop a systematists are more likely to view species within the taxonomic system based on criteria identified by context of clades (i.e. monophyletic groups of species) observers. For instance, organisms could be clustered and ancestor-descendant relationships. In this on the basis of similarity (Sneath and Sokal, 1973) or historical perspective, a species can be an explicitly on the basis of shared characters (Nelson, 1989) historical entity (e.g. Simpson’s evolutionary species), without assumptions of underlying biological or a contemporaneous entity that is defined by processes. A taxonomic species identified in this way historical relationships to others (e.g. (Cracraft, 1989) would be a class of organisms, and would have no version of a phylogenetic species). Under the genetic necessary relationship to genetic species that occur as species concept, a species is a group of organisms that individuals in nature. share genetic drift. This is a contemporaneous CONCLUSIONS definition. The genetic species concept within the context of species concept debates In this paper, the species problem has been identified with two ongoing scientific debates: first, over the meaning of the word "species"; and second, over methods of species identification. In practice, many disputes take specific forms that do not fall into just one of these categories of debate. Endler (1989) has outlined four dichotomies, each a broad category of species concept debate, and it may be informative to discuss the genetic species concept within the context of each of these. Taxonomic versus evolutionary Endler (1989) used “taxonomic” to refer to those species concepts narrow sense, the genetic species concept is not a motivated purely by the need for classification. The ‘reproductive’ concept because it does not apply only merit of a taxonomic concept of species is determined to sexual organisms. However, the reproductive by how well it works for identification and grouping, successes of different organisms are linked within a without regard to the evolutionary relatedness of genetic species, and this occurs regardless of sex. The organisms, or the processes that gave rise to species. genetic species concept does fit the ‘cohesive’ category Endler’s category of “evolutionary” applies to in some respects. Genetic drift has long been concepts that are defined in terms of, or make recognized as a force that causes homogeneity among assumptions about, evolutionary processes. The groups of organisms that share drift, and genetic drift genetic species concept is motivated solely by process, is a significant component of Templeton’s cohesion and thus would fall squarely in the “evolutionary” species concept (Templeton, 1989). However, under category. the genetic species concept, it is possible that a genetic Theoretical vs operational A theoretical species can form by the admixture of organisms that concept of species is concerned with the meaning of the word “species” in terms of the basic processes that give rise to species. Theoretical concepts, like the biological and the genetic species concepts, may not be very practical for purposes of species identification. Operational species concepts may include taxonomic concepts (see 1, above), because of the focus on applicability, but they also include species concepts The genetic species concept seems to be a that have a dual focus on evolutionary processes and reduction of those other species concepts that focus on practicality. Perhaps the best examples of these are the biological processes. The terms of the genetic species various attempts to develop a phylogenetic species concept (e.g. “DNA replication” and “genetic drift”) concept (Rosen, 1979; Donoghue, 1985; Cracraft, refer to things that are components and properties of 1989; Baum, 1992; Davis and Nixon, 1992). the things in other species concepts. The genetic Contemporaneous vs clade Many species concept is not at odds with other concepts that evolutionary biologists, including population biologists, focus on extant organisms, and rely on meanings of “species” that hinge on ongoing biological processes. The biological and the ecological Reproductive vs Cohesive Endler (1989) classifies as “reproductive”, those concepts that focus on the reproductive processes that maintain either isolation between species or recognition within species. He contrasts these with concepts that focus on species as units with genetic and phenotypic cohesion. The distinction between these, and thus the failure of some species concepts, must be recognized because of numerous examples of groups of organisms that appear to be one reproductive species because of the exchange of genes, but yet persist as multiple distinct cohesive units. The genetic species concept does not simply fit into either of Endler’s categories. In a are not from genetic species or are from multiple genetic species. Such an admixture may exist as a group of organisms that share genetic drift, but at the time of initial mixing the group may be less cohesive than any of the groups from which the mixture was formed. The meaning of “reduction” are described in terms of biological processes. Other concepts are extensions of the genetic species concept in different contexts (e.g. ecological species, biological species, and evolutionary species). Components of J. Hey Genetic Species 12 these concepts are also included within the cohesion The process of species identification is not species concept (Templeton, 1989). A cohesion simplified by having a meaning of "species". Instead species is "an evolutionary lineage that serves as the the task is shown to be as difficult as measuring arena of action of basic microevolutionary forces, such genetic drift and then deciding if the data contain as gene flow (when applicable), genetic drift and evidence of partitions in the structure of genetic drift. natural selection" (Templeton, 1994). Thus the It may not be possible to identify species in many cohesion concept resembles the evolutionary species cases, but it is clearer why species can be so difficult to concept with a population genetic emphasis on the identify. causes of phenotypic similarity within species. Explaining the existence of species However all of the basic microevolutionary forces can only occur when organisms share, to some degree, in a birth and death process and thus in the process of genetic drift. The reduction of “species” has many parallels with the reduction of “gene”. Beginning with Johannsen’s strictly operational definition and extending to the time of the elucidation of the function of DNA sequences, the nature of genes and the meaning of “gene” were at the heart of great debate and energetic research (Dunn, 1965). Today it is clear that the function of DNA sequences takes many forms. For example, an incomplete list includes protein coding sequences, ribosomal RNA coding sequences, transfer RNA coding sequences, introns, promoters, enhancers, and telomeric sequences. Furthermore, the boundaries for some of these functional sequences, in terms of their location in a DNA chain, may not be distinct; and there are many cases where DNA sequences associated with one function are part of a larger sequence that is associated with a different function (e.g. enhancers lying within protein coding regions). Johannsen’s single word is still useful as an approximation in many cases, but it does not begin to represent the breadth of functional diversity manifest in DNA sequences. In the day to day practice of molecular genetic research, the word “gene” has largely been replaced by a richer lexicon that has come with the growth of understanding of DNA function. Like ‘gene’, the word ‘species’ may not be necessary or always useful in a research context, particularly when situations occur that do not fit a definition. However, understanding the meaning of “species” does simplify a plan for inquiry on the causes of species diversity. Speciation research should be the science of genetic history and of the biotic and abiotic phenomena that determine whether or not organisms share genetic drift and that shape the gene tree histories of closely related organisms. It is not necessary that the species status of organisms be established prior to research. If an assessment of the history of genetic drift for a sample of organisms reveals a relatively simple picture of one or more groups that share genetic drift, then species status may be delineated. However application of species status and species distinctions will not add information or meaning. In practice, an investigation into the nature of species often begins with an implicit or explicit viewpoint that species exist. The approach taken in this paper has been to assume the existence of some relatively simple natural phenomena, and then to consider whether these things will cause something that corresponds to other ideas on species. In particular, the starting point includes the existence of replicating DNAs, an assumption that mutations can occur, and that the real world has limited resources and imposes geographic constraints on the location of DNAs. With this small set of observations it has been possible to predict several things about the diversity of DNAs that seem to correspond to common observations about the diversity of organisms and that are often attributed to species. 1) Mutations and geography will have the effect of partitioning DNAs into different groups, within each of which there is a shared process of genetic drift caused by the random replacement of some DNAs by the descendants of others. 2) Mutation will in the long run have a greater effect than geography on the formation of groups. Mutation is, for the most part, not reversible (see , THE CAUSES OF GENETIC SPECIES) while geography can change and can be overcome by the movement of DNAs. 3) There will often occur situations where the pattern of genetic drift does not occur with distinct boundaries among groups of DNAs, but rather occurs at varying rates across a geographic range. In these situations, the group status of DNAs will appear ambiguous. 4) DNAs that share a process of genetic drift will have a gene tree history shaped by that process. If they and their descendants persist in sharing genetic drift, then this gene tree will become monophyletic. 5) Groups of DNAs that share genetic drift can occur whether or not the DNAs exchange portions of their sequence (i.e. engage in sex). In short, a very small set of basic observations lead to predictions that correspond very well to many of our most basic observations about groups of organisms, and that are often attributed to species. In this sense, the existence of species has been at least partly explained. Also, since the basic observations J. Hey Genetic Species 13 (i.e. replicating DNA, mutations, geography and Sunderland,Mass. limiting resources) are not in doubt and are relatively Davis, J. I., and K. C. Nixon. 1992. Populations, simplistic (i.e. being chemical and physical, but not genetic variation, and the delimitation of very biological), these observations seem to be phylogenetic species. Syst. Biol. 41:421-435. necessary for any complete explanation of species. It Dawkins, R. 1978. Replicator selection and the is then appropriate to ask whether these observations extended phenotype. Zeitschrift fur Tierpsychol are a sufficient explanation of species. There remain 47:61-76. two distinct contexts in which the genetic species de Queiroz, K., and M. J. Donoghue. 1988. concept may be judged incorrect by reason of Phylogenetic systematics and the species insufficiency. First, the genetic species concept may problem. Cladistics 4:317-338. imply some unforeseen consequence that is not a true Dobzhansky, T. 1950. Mendelian populations and their phenomenon. As described, genetic drift is an evolution. Amer. Natur. 84:401-418. epiphenomenon of the basic components of replicating ---. 1951. Genetics and the origin of species, 3rd ed. DNA, mutations, geography and limiting resources. If Columbia University Press, New York. it were shown that the simple model that has been Donoghue, M. J. 1985. A critique of the biological constructed also leads to predictions that do not occur species concept and recommendations for a in nature, then the model is incomplete and something phylogenetic alternative. The Bryologist must be added. Second, something may have been 88:172-181. overlooked in the true nature of species. It may be that Dunn, L. C. 1965. A short history of genetics. all real species share a property that cannot be McGraw-Hill Book Company, New York. predicted by the simple model developed here.Endler, J. A. 1989. Conceptual and other problems inACKNOWLEDGMENTSspeciation, pp. 625-648. In D. Otte and J. A.I thank David Baum, Michael Bulmer, RobDeSalle, Adam Eyre-walker, Douglas Futuyma,Michael Ghiselin, Anna Graybeal, Brandon Gaut,Holly Hilton, David Houle, Richard Kliman, NicholasLeahy, Richard Lewontin, Margaret Saks, Elliot Sober,Alan Templeton, Robert Trivers, David Wake, JohnWakeley, and Rong Lin Wang for comments on themanuscript. This research was supported by a grantfrom NSF (DEB-9306625). REFERENCESAvise, J. C., and R. M. J. Ball. 1990. Gene genealogiesand the coalescent process, pp. 45-67. In P. H. Harvey and L. Partridge (eds.), Oxford surveysset of external conditions. J. Linn. Soc. Lond.in evolutionary biology, vol. 7, OxfordZool. 11:496-505.University Press, New York.---. 1888. Divergent evolution through cumulativeBaum, D. 1992. Phylogenetic species concepts. Tr.segregation. J. Linn. Soc. Lond. Zool. 20:189-Ecol. Evol. 7:1-2.274.Baum, D. A., and K. L. Shaw. 1995. GenealogicalHartl, D. L., and D. E. Dykhuizen. 1984. Theperspectives on the species problem, pp. 289-population genetics of Escherichia coli. Annual303. In P. C. Hock and A. G. Stevenson (eds.),Review of Genetics 18:31-68. Experimental and molecular approaches to plantbiosystematics, Missouri Botanical Garden, St.Louis.Charlesworth, B., M. T. Morgan, and D. Charlesworth.1993. The effect of deleterious mutations onneutral molecular evolution. Genetics 134:1289-1303.Cracraft, J. 1989. Speciation and its ontology: theempirical consequences of alternative speciesconcepts for understanding patterns andprocesses of differentiation, p. mybook. In D. Otte and J. A. Endler (eds.), Speciation and itsconsequences, Sinauer Associates,Endler (eds.), Speciation and its consequences,Sinauer Associates, Sunderland, Mass.Felsenstein, J. 1974. The evolutionary advantage ofrecombination. Genetics 78:737-756.Ghiselin, M. T. 1966. On psychologism in the logic oftaxonomic controversies. Syst. Zool. 15:207-215.---. 1974. A radical solution to the species problem. Syst. Zool. 23:536-544.---. 1987. Species concept, individuality, andobjectivity. Biology and Philosophy 2:127-143.Gilbert, W. 1986. The RNA world. Nature 319:618.Gulick, J. T. 1872. On diversity of evolution under one Hennig, W. 1979a. Phylogenetic systematics.University of Illinois Press, Chicago, 73.---. 1979b. Phylogenetic systematics. University ofIllinois Press, Chicago, 59.---. 1979c. Phylogenetic systematics. University ofIllinois Press, Chicago, 73.---. 1979d. Phylogenetic systematics. University ofIllinois Press, Chicago, 81.Hey, J. 1991. A multi-dimensional coalescent processapplied to multi-allelic selection models andmigration models. Theor. Pop. Biol. 39:30-48.---. 1994. Bridging phylogenetics and populationgenetics with gene tree models. In B. J. Hey Genetic Species 14Schierwater, B. Streit, G. Wagner and R.358:203-209.DeSalle (eds.), Molecular Approaches toRosen, D. E. 1979. Fishes from the uplands andEcology and Evolution, Birkhäuser-intermontane basins of Guatemala: revisionaryVerlag, Basel.studies and comparative biogeography. Bull.Hey, J., and R. M. Kliman. 1993. Population geneticsAm. Mus. Nat. Hist. 162:267-376.and phylogenetics of DNA sequence variation at Simpson, G. G. 1961. Principles of animal taxonomy.multiple loci within the DrosophilaColumbia University Press, New York, 150-melanogaster species complex. Mol. Biol. Evol.153.10:804-822.Slatkin, M. 1987. The average number of sitesHill, W. G., and A. Robertson. 1966. The effect ofseparating DNA sequences drawn from alinkage on limits to artificial selection. Genet.subdivided population. Theor. Pop. Biol. 32:42-Res. Camb. 8:269-294.49.Hudson, R. R. 1990. Gene genealogies and theSneath, P. H. A., and R. R. Sokal. 1973. Numericalcoalescent process, pp. 1-44. In P. H. HarveyTaxonomy. Freeman, San Francisco. and L. Partridge (eds.), Oxford Surveys inEvolutionary Biology, vol. 7, Oxford UniversityPress, New York.Hudson, R. R., and N. L. Kaplan. 1988. The coalescentprocess in models with selection andrecombination. Genetics 120:819-829.Hull, D. L. 1976. Are species really individuals. Syst.Zool. 15:174-191.---. 1978. A matter of individuality. Philosophy ofScience 45:335-360. Hutchinson, G. E. 1958. Concluding remarks. ColdSpring Harbor Symposium in QuantitativeBiology 22:415-427.Kaplan, N., R. R. Hudson, and C. H. Langley. 1989.The"hitchhiking effect"revisited. Genetics123:887-899.J. A. Endler (eds.), Speciation and itsKingman, J. F. C. 1982. The coalescent. Stochasticconsequences, Sinauer Associates,Process. Appl. 13:235-248.Sunderland,Mass.Levin, D. A. 1979. The nature of plant species. Science ---. 1994. The role of molecular genetics in speciation204:381-384.studies. In B. Schierwater, B. Streit, G. WagnerLewontin, R. C. 1970. The units of selection. AnnualReview of Ecology and Systematics 1:1-18.Masters, J. C., and H. G. Spencer. 1989. Why we needa new genetic species concept. Syst. Zool.38:270-279.Maynard Smith, J., and J. Haigh. 1974. The hitch-hiking effect of a favourable gene. Genet. Res.Camb. 23:23-35.Mayr, E. 1942. Systematics and the origin of species.Columbia University Press, New York.---. 1963. Populations, species and evolution. HarvardUniversity Press, Cambridge, Mass.---. 1982. The growth of biological thought. HarvardUniversity Press, Cambridge, MA. Nelson, G.1989. Species and taxa: systematics andevolution, pp. 60-81. In D. Otte and J. A. Endler (eds.), Speciation and its consequences, SinauerAssociates, Sunderland, Mass.Nixon, K. C., and Q. D. Wheeler. 1990. Anamplification of the phylogenetic speciesconcept. Cladistics 6:211-223.Orgel, L. E. 1992. Molecular replication. NatureTajima, F. 1989. The effect of change in populationsize on DNA polymorphism. Genetics 123:597-601.Takahata, N. 1988. The coalescent in two partiallyisolated diffusion populations. Genet. Res.Camb. 52:213-222.Tavaré, S. 1984. Line-of-Descent and genealogicalprocesses, and their applications in populationgenetics models. Theor. Pop. Biol. 26:119-164.Templeton, A. R. 1981. Mechanisms of speciationapopulation genetic approach. Annual Review ofEcology and Systematics 12:23-48.---. 1989. The meaning of species and speciation: agenetic perspective",b"Speciation and itsconsequences, p. xrxandmybook. In D. Otte and and R. DeSalle (eds.), Molecular Approaches toEcology and Evolution, Birkhäuser-Verlag,Basel.Van Valen, L. 1976. Ecological species, multispecies,and oaks. Taxon 25:233-239.Vrana, P., and W. Wheeler. 1992. Individualorganisms as terminal entities: laying the speciesproblem to rest. Cladistics 8:67-72.Watterson, G. A. 1982. Mutant substitutions at linkednucleotide sites. Adv. Appl. Prob. 14:206-224.Wright, S. 1931. Evolution in mendelian populations.Genetics 16:97-159.---. 1943. Isolation by distance. Genetics 28:114-138.